“The intuitive machine learning component of the system is anticipated to reduce electricity and natural gas use throughout the building by constantly adjusting set-points as occupancy, solar irradiance, and internal heat loads vary throughout the day. ”
— Triovest Realty Advisors
Project Innovation
Intelligent design for better building operation
As tenant expectations evolve and sustainable operations become the norm, the problem of optimizing HVAC systems has become increasingly complex. Conventionally, Variable Air Volume (VAV) systems are configured to meet a specific static pressure in the ducting system – this is defined during installation and may be revisited many years later. The intent is that fan speed will be controlled to maintain this static pressure and provided the static pressure set point is met, full airflow will be provided to all VAV boxes to meet the maximum expected or potential demand. The issue is that this is almost never the case in practice and, as a result, extra fan energy is wasted not to supply airflow but to add pressure into the ducting system. Existing control improvements to reset static pressure can be effective but they can be challenging to program or rendered ineffective by malfunctioning zone with high demand. On top of
this, most VAV systems have a constant supply air temperature or a simple reset strategy which does not factor in the total system energy. The ability to intelligently optimize the HVAC system with an artificial intelligence (AI) powered strategy instead of the traditional trial-and-error approach, can be done using Dynamic VAV Optimization (DVO), an HVAC optimization strategy that relies on a cloud-based, machine learning powered algorithm to control Air Handling Unit (AHU) fan speed, supply temperatures and humidity levels. These are two key elements of VAV systems which DVO seeks to improve on – a more dynamic static pressure reset and holistic control of heating, cooling, and fan energy to maximize system efficiency.
The key metrics for success from the application of this technology will be energy savings and resulting greenhouse gas (GHG) emission reductions achieved compared to both baseline operation as well as existing energy efficiency strategies for VAV HVAC systems, while continuing to meet the comfort condition targets. The intuitive machine learning component of the system is anticipated to reduce electricity and natural gas use throughout the building by constantly adjusting set-points as occupancy, solar irradiance, and internal heat loads vary throughout the day.
Project Findings
Visit the Resource Library to find resources on this project and others like it (search popular subjects: ‘Automation & Smart Technology’).
Project Summary Pages
